Genome-wide changes in DNase-hypersensitivity during osteoblastogenesis reveal differential usage of DNA motifs and define novel cis-regulatory regions that control gene expression during differentiation [DNase-Seq]

Gene regulation during the process of osteoblastogenesis has been well-described, yet the discovery of novel regulatory regions has been limited by how we currently predict the locations of functional cis-regulatory modules. Historically, the de novo identification of sequences critical for the control of gene expression relied primarily on sequence conservation in promoters. Queries for binding motifs were based on position weight matrices of known transcription factors and the identification of disease-causing, non-coding mutations near critical genes. However, we now must consider that regulatory elements also rely on 3-dimensional chromosomal interactions between far-distal regions, epigenetic chromosomal modifications, and RNA:DNA interactions. Traditionally, DNaseI-hypersensitivity assays have been used for the identification of regulatory regions via preferential digestion at chromatin depleted or displaced of nucleosomes, as a result of transcription factor occupancy. We probed DNase hypersensitivity on a genome-wide scale to determine whether osteogenic differentiation and/or bone-related gene regulation is marked by the presence of commonly utilized DNA motifs within active cis-regulatory modules. We thus sought to evaluate the gain or loss of motif representation within hypersensitive regions during osteoblastogenesis, from day-0 (growth-phase) to day-28 (mineralizing) MC3T3 cultures. We find that differentiation is marked by an increased enrichment of NFkB-p65, MEF2, and bHLH/E-box motifs within hypersensitive regions, while CTCF, NF1, TEAD, and AP1 motifs decrease. Furthermore, grouping hypersensitive regions based on genomic positioning (promoters, introns, exons, and far-distal regions) reveals significant differences in motif abundance in first introns versus other genomic positions. This finding suggests that the regulation conferred within first intron sequences may be somewhat distinct. Interestingly, the majority of motifs that were enriched, regardless of genomic position or differentiation time-point, were not completely matched to currently known transcription factor motifs (curated in the JASPAR database). Taken together, the changes in DNase-hypersensitive regions during osteoblastogenesis and the enrichment of distinct motifs within these regions indicate that osteoblasts utilize unique sets of motif rules for transcription factor binding or that regulatory control operates through undiscovered factors. Genome-wide DNase hypersensitivity mapping of osteoblast cultures was performed by adapting the DNase-seq protocol from Song et al. (Song and Crawford, 2010) with slight modifications. Growth-phase (day 0), matrix-deposition stage (day 9), or mineralization stage (day 28) MC3T3-E1 clone-4 cultures were subjected to DNase-seq library preparation. Libraries of purified DNA were generated using custom adapters described in Song et al. High-throughput sequencing was performed by Illumina Genome Analyzer II with 36 base reads and on an Illumina Hiseq-1000 with 100 base reads. Base calls and sequence reads were generated by Illumina CASAVA software (version 1.6, Illumina). Two independent biological repeats of DNase-seq libraries were prepared for each time point. Each biological repeat is represented by two technical repeats.